Optimizing storage and memory systems for energy and performance

Electrical energy is an essential resource for the modern world. Unfortunately, its price has almost doubled in the last decade. Furthermore, energy production is also currently one of the primary sources of pollution. These concerns are becoming more important in data-centers. As more computational power is required to serve hundreds of millions of users, bigger data-centers are becoming necessary. This results in higher electrical energy consumption. Of all the energy used in data-centers, including power distribution units, lights, and cooling, computer hardware consumes as much as 80%. Consequently, there is opportunity to make data-centers more energy efficient by designing systems with lower energy footprint. Consuming less energy is critical not only in data-centers. It is also important in mobile devices where battery-based energy is a scarce resource. Reducing the energy consumption of these devices will allow them to last longer and re-charge less frequently. Saving energy in computer systems is a challenging problem. Improving a system's energy efficiency usually comes at the cost of compromises in other areas such as performance or reliability. In the case of secondary storage, for example, spinning-down the disks to save energy can incur high latencies if they are accessed while in this state. The challenge is to be able to increase the energy efficiency while keeping the system as reliable and responsive as before. This thesis tackles the problem of improving energy efficiency in existing systems while reducing the impact on performance. First, we propose a new technique to achieve fine grained energy proportionality in multi-disk systems; Second, we design and implement an energy-efficient cache system using flash memory that increases disk idleness to save energy; Finally, we identify and explore solutions for the page fetch-before-update problem in caching systems that can: (a) control better I/O traffic to secondary storage and (b) provide critical performance improvement for energy efficient systems.

Interoperable resource brokering with policy-based provisioning and job allocation

The increasing needs for computational power in areas such as weather simulation, genomics or Internet applications have led to sharing of geographically distributed and heterogeneous resources from commercial data centers and scientific institutions. Research in the areas of utility, grid and cloud computing, together with improvements in network and hardware virtualization has resulted in methods to locate and use resources to rapidly provision virtual environments in a flexible manner, while lowering costs for consumers and providers. ^ However, there is still a lack of methodologies to enable efficient and seamless sharing of resources among institutions. In this work, we concentrate in the problem of executing parallel scientific applications across distributed resources belonging to separate organizations. Our approach can be divided in three main points. First, we define and implement an interoperable grid protocol to distribute job workloads among partners with different middleware and execution resources. Second, we research and implement different policies for virtual resource provisioning and job-to-resource allocation, taking advantage of their cooperation to improve execution cost and performance. Third, we explore the consequences of on-demand provisioning and allocation in the problem of site-selection for the execution of parallel workloads, and propose new strategies to reduce job slowdown and overall cost.^

Interoperable Resource Brokering with Policy-based Provisioning and Job Allocation

The increasing needs for computational power in areas such as weather simulation, genomics or Internet applications have led to sharing of geographically distributed and heterogeneous resources from commercial data centers and scientific institutions. Research in the areas of utility, grid and cloud computing, together with improvements in network and hardware virtualization has resulted in methods to locate and use resources to rapidly provision virtual environments in a flexible manner, while lowering costs for consumers and providers. However, there is still a lack of methodologies to enable efficient and seamless sharing of resources among institutions. In this work, we concentrate in the problem of executing parallel scientific applications across distributed resources belonging to separate organizations. Our approach can be divided in three main points. First, we define and implement an interoperable grid protocol to distribute job workloads among partners with different middleware and execution resources. Second, we research and implement different policies for virtual resource provisioning and job-to-resource allocation, taking advantage of their cooperation to improve execution cost and performance. Third, we explore the consequences of on-demand provisioning and allocation in the problem of site-selection for the execution of parallel workloads, and propose new strategies to reduce job slowdown and overall cost.

Physical dynamic simulation of shipboard power system components in a distributed computational environment

Shipboard power systems have different characteristics than the utility power systems. In the Shipboard power system it is crucial that the systems and equipment work at their peak performance levels. One of the most demanding aspects for simulations of the Shipboard Power Systems is to connect the device under test to a real-time simulated dynamic equivalent and in an environment with actual hardware in the Loop (HIL). The real time simulations can be achieved by using multi-distributed modeling concept, in which the global system model is distributed over several processors through a communication link. The advantage of this approach is that it permits the gradual change from pure simulation to actual application. In order to perform system studies in such an environment physical phase variable models of different components of the shipboard power system were developed using operational parameters obtained from finite element (FE) analysis. These models were developed for two types of studies low and high frequency studies. Low frequency studies are used to examine the shipboard power systems behavior under load switching, and faults. High-frequency studies were used to predict abnormal conditions due to overvoltage, and components harmonic behavior. Different experiments were conducted to validate the developed models. The Simulation and experiment results show excellent agreement. The shipboard power systems components behavior under internal faults was investigated using FE analysis. This developed technique is very curial in the Shipboard power systems faults detection due to the lack of comprehensive fault test databases. A wavelet based methodology for feature extraction of the shipboard power systems current signals was developed for harmonic and fault diagnosis studies. This modeling methodology can be utilized to evaluate and predicate the NPS components future behavior in the design stage which will reduce the development cycles, cut overall cost, prevent failures, and test each subsystem exhaustively before integrating it into the system.

Physical dynamic simulation of shipboard power system components in a distributed computational environment

Shipboard power systems have different characteristics than the utility power systems. In the Shipboard power system it is crucial that the systems and equipment work at their peak performance levels. One of the most demanding aspects for simulations of the Shipboard Power Systems is to connect the device under test to a real-time simulated dynamic equivalent and in an environment with actual hardware in the Loop (HIL). The real time simulations can be achieved by using multi-distributed modeling concept, in which the global system model is distributed over several processors through a communication link. The advantage of this approach is that it permits the gradual change from pure simulation to actual application. In order to perform system studies in such an environment physical phase variable models of different components of the shipboard power system were developed using operational parameters obtained from finite element (FE) analysis. These models were developed for two types of studies low and high frequency studies. Low frequency studies are used to examine the shipboard power systems behavior under load switching, and faults. High-frequency studies were used to predict abnormal conditions due to overvoltage, and components harmonic behavior. Different experiments were conducted to validate the developed models. The Simulation and experiment results show excellent agreement. The shipboard power systems components behavior under internal faults was investigated using FE analysis. This developed technique is very curial in the Shipboard power systems faults detection due to the lack of comprehensive fault test databases. A wavelet based methodology for feature extraction of the shipboard power systems current signals was developed for harmonic and fault diagnosis studies. This modeling methodology can be utilized to evaluate and predicate the NPS components future behavior in the design stage which will reduce the development cycles, cut overall cost, prevent failures, and test each subsystem exhaustively before integrating it into the system.

Computational modeling of intermediate temperature proton exchange membrane (PEM) fuel cells

A two-phase three-dimensional computational model of an intermediate temperature (120--190°C) proton exchange membrane (PEM) fuel cell is presented. This represents the first attempt to model PEM fuel cells employing intermediate temperature membranes, in this case, phosphoric acid doped polybenzimidazole (PBI). To date, mathematical modeling of PEM fuel cells has been restricted to low temperature operation, especially to those employing Nafion ® membranes; while research on PBI as an intermediate temperature membrane has been solely at the experimental level. This work is an advancement in the state of the art of both these fields of research. With a growing trend toward higher temperature operation of PEM fuel cells, mathematical modeling of such systems is necessary to help hasten the development of the technology and highlight areas where research should be focused.^ This mathematical model accounted for all the major transport and polarization processes occurring inside the fuel cell, including the two phase phenomenon of gas dissolution in the polymer electrolyte. Results were presented for polarization performance, flux distributions, concentration variations in both the gaseous and aqueous phases, and temperature variations for various heat management strategies. The model predictions matched well with published experimental data, and were self-consistent.^ The major finding of this research was that, due to the transport limitations imposed by the use of phosphoric acid as a doping agent, namely low solubility and diffusivity of dissolved gases and anion adsorption onto catalyst sites, the catalyst utilization is very low (∼1--2%). Significant cost savings were predicted with the use of advanced catalyst deposition techniques that would greatly reduce the eventual thickness of the catalyst layer, and subsequently improve catalyst utilization. The model also predicted that an increase in power output in the order of 50% is expected if alternative doping agents to phosphoric acid can be found, which afford better transport properties of dissolved gases, reduced anion adsorption onto catalyst sites, and which maintain stability and conductive properties at elevated temperatures.^